作者: Wijbrandt H. van Schuur
DOI: 10.1016/S0166-4115(08)60240-X
关键词: Dimension (data warehouse) 、 Preference (economics) 、 Representation (mathematics) 、 Multidimensional analysis 、 Rank (computer programming) 、 Social psychology 、 Scale (descriptive set theory) 、 German 、 Psychology 、 Mathematical economics 、 Survey data collection
摘要: This paper discusses a number of problems with existing unfolding models and proposes strategy analysis to overcome these problems. assumes dichotomous or dichotomized data, derives unfoldability criteria from information about ordered triples stimuli. A unidimensional scale conforming can be found for maximal subset procedure applied full partial rank orders preference, which are “pick k/n” Likert-type rating scales, any/n” data. is applicable large data sets, such as survey As an example, the preferences five German political parties in electoral surveys 1969, 1972, 1980. dominant left-right dimension found, violations this representation discussed.